Opportunities for AI in Health Equity, Medical Education, and Beyond
Expert Perspectives · Waqas Haque, MD, MPH · 2024-10-25
October’s edition of Dr. Waqas Haque’s blog series on artificial intelligence (AI) and technology features new uses of AI in providing prognostic information to guide treatment decisions in prostate cancer, analyzing crowdfunding stories on GoFundMe to understand financial hardships and social determinants of health, and integrating large language models into medical education for radiation oncology.
Leveraging Biomarkers With AI
Our upcoming AI and technology podcast episode in November will feature Dr. Trevor Royce, a radiation oncologist with multiple industry experiences in using AI and machine learning to drive cancer care. He co-authored an article earlier this month validating a multimodal artificial intelligence (MMAI) prognostic biomarker (known as Artera AI Prostate) for high-risk prostate cancer using data from six phase 3 randomized trials (n=1,088 men with at least one NCCN high/very-high risk factor).
The MMAI model was independently prognostic for distant metastasis, prostate cancer–specific mortality, and death with distant metastasis, even after adjusting for clinical risk factors. Patients in MMAI quartiles 3-4 had significantly higher 10-year risk of distant metastasis (26%) compared to quartiles 1-2 (8%). Importantly, the biomarker identified highly variable risks within the high-risk patient group. Another advantage of the selected sample is that 20% of patients were Black. This scalable MMAI tool can provide more personalized prognostic information to guide treatment decisions for high-risk prostate cancer patients. Future research can evaluate how the biomarker interacts with specific treatments.
Addressing Health Inequity With AI
A recent research latter in JAMA Oncology demonstrates an innovative application of artificial intelligence in healthcare research, specifically for investigating social determinants of health among cancer survivors. The study team utilized OpenAI’s ChatGPT (version 3.5) to analyze over 24 million words from nearly 90,000 cancer-related crowdfunding stories on GoFundMe. This large-scale natural language processing (NLP) approach allowed for efficient extraction of quantitative data from narrative text, overcoming the logistical challenges of manual analysis. The AI model was programmed to interpret each fundraiser story and generate outputs based on predefined questions about patient demographics, cancer diagnosis, treatments, and reasons for financial assistance.
The analysis revealed that 25.5% of stories mentioned medical financial hardship, while 24.1% described health-related social needs such as employment disruptions, income loss, and struggles with basic needs like food and housing. Most fundraisers (88.5%) requested $5,000 or more, with 43.4% seeking $15,000 or more. While 99.5% of campaigns received some funds, only 11.4% reached their goals. The use of AI in this context demonstrates its potential for processing large volumes of unstructured text to gain insights into social determinants of health for patients with cancer and the need for stronger safety nets.
AI for Medical Education
A study in AI in Precision Oncology evaluated the performance of seven large language models (LLMs) on the 2021 American College of Radiology’s Radiation Oncology In-Training Examination. OpenAI’s GPT-4-turbo performed best, answering 74.2% of questions correctly, comparable to upper-level radiation oncology residents. While LLMs excelled in statistics, they generally underperformed in clinical areas (except for GPT-4-turbo). Performance varied across disease sites, with LLMs scoring lowest in gastrointestinal, genitourinary, and gynecology domains. The study highlights the potential of AI in medical education in radiation oncology. Like other studies that we have highlighted, a recurring theme is the need for expert oversight and further validation to ensure accuracy in specialized medical fields.
About Dr. Haque
Waqas Haque, MD, MPH, is a Hematology/Oncology Fellow at the University of Chicago. He recently completed his Internal Medicine Residency at New York University (NYU) in a Clinical Investigator Track. Dr. Haque has balanced his patient care work with a variety of research projects. During his fellowship training at the University of Chicago, he plans to further his work in innovative clinical trial design, value-based care delivery to cancer patients, and clinical investigation.